NEMT by the numbers: riders, trips, spending, and drivers (2026)

Good NEMT numbers come from a handful of federal sources, and each one sees only part of the picture. CMS counts Medicaid riders from claims data, MACPAC estimated spending, the Census Bureau counts businesses with payroll, BLS counts wage earners, and the HHS inspector general tallies fraud cases. The figures below come from those sources, each with its year. Where a number has a known blind spot, it is noted next to the number.

The headline numbers

The rest of this page explains where each number comes from and what it leaves out.

Riders and ride days

CMS’s expanded report to Congress, dated June 20, 2023, is the newest national count of Medicaid NEMT use. It measures ride days, meaning the number of separate days a member had at least one ride. A round trip to dialysis and back counts as one ride day, so ride days understate the number of trips a company actually runs.

The pandemic explains most of the drop. Between February and April 2020, monthly ride days fell 65 percent as visits moved to telehealth or were put off. Ride days in 2021 still averaged about 30 percent below pre-pandemic months. CMS treats every figure as a floor, since states that pay rides as an administrative expense may not report them in the claims data, and Puerto Rico, Utah, and the U.S. Virgin Islands were left out of the 2021 national totals.

MACPAC’s separate analysis for fiscal 2018 lands in the same range: about 3.2 million users, 4.8 percent of beneficiaries, and 61.5 million ride days.

Need runs wider than use. A study of National Health Interview Survey data, published in the American Journal of Public Health in 2020, estimated that 5.8 million people in the United States (1.8 percent) put off medical care in 2017 because they had no way to get there. Medicaid recipients and people with a functional limitation had higher odds of reporting that barrier.

Enrollment sets the size of the pool. CMS’s June 2026 data counts 65,969,381 people in Medicaid. In its June 3, 2026 rule on the new community engagement requirement, CMS projects enrollment 2.3 million lower in fiscal 2027 and 3.1 million to 3.3 million lower in later years. Our guide to Medicaid work requirements and NEMT covers what that means for trip volume.

Who rides the most

Heavy users are a small share of members but a large share of trips. CMS’s 2021 figures by group:

Members with kidney failure made up less than half a percent of all Medicaid members in 2021, yet nearly half of them used rides, on more than five days a month. That pattern is why dialysis transportation contracts are the backbone of many fleets.

Where members live matters too. In frontier and remote areas, about 7 percent of members used NEMT in 2021. Across states, the share of members riding ranged from 1 percent (Maryland) to almost 11 percent (Alaska), and ride days per user ran from 0.3 a month in Alabama and Maryland to 3.7 in Wisconsin. Part of that spread reflects how each state pays for and reports rides, which how states run NEMT explains.

What the rides are for

Routine care takes the largest share of ride days. In each year from 2018 to 2021, preventive visits accounted for 29 to 31 percent of ride days, visits for acute and chronic conditions for 20 to 21 percent, and lab and imaging work for 16 to 19 percent. A few recurring treatments carry far more weight than their rider counts suggest.

• Dialysis. Between 8 and 13 percent of all ride days went to dialysis, depending on the year, even though only 0.1 percent of members used NEMT for it. Those riders averaged 4.4 to 4.8 ride days a month, the most of any service type. Dialysis accounted for 68 percent of ride days among members with kidney failure in 2021.
• Behavioral health. Rides to behavioral health services made up 18 to 19 percent of ride days before the pandemic and 16 percent during it.
• Substance use treatment. All substance use treatment took 11 to 14 percent of ride days. Medication treatment for opioid use disorder alone took 5 to 8 percent, with riders averaging 3 to 3.5 ride days a month. See opioid treatment program transportation.

Services that need an in-person visit gained share during the pandemic. Dialysis rose from 8 percent of ride days in 2019 to 13 percent in 2020, the largest shift of any service type.

How riders travel

Taxis and vans carry most Medicaid trips. In the 2021 data, CMS assigned each ride to a mode from its billing code, or from the provider’s taxonomy code when the billing code did not say.

The mix shifts with geography. In large central metro areas, taxis carried 42.6 percent of ride days and vans 29.0 percent. In frontier and remote areas, vans carried 41.3 percent, and private vehicles rose to 17.1 percent. MACPAC’s fiscal 2018 data, which drops ambulances from its definition, found vans at 46 percent of ride days and taxis at 36.7 percent.

Spending

No federal source publishes a complete, current total for NEMT spending. MACPAC’s estimate is the most complete: $2.6 billion in state and federal Medicaid spending in fiscal 2018, about $40 per full-year-equivalent enrollee.

• Who got paid. About $1.7 billion was paid by states directly or through fee-for-service broker arrangements. About $0.9 billion went to prepaid transportation brokers.
• What it leaves out. Payments that managed care plans made to transportation providers are not in the figure, and neither are ambulances. The real total is higher.
• How trips were paid. Of the 61.5 million ride days in fiscal 2018, about 23 million were paid fee-for-service. The other two-thirds were paid under capitated broker or health plan arrangements.

Delivery models explain much of that split. MACPAC counted 35 states using a transportation broker for at least some members or areas as of June 2021, and 26 states that had written NEMT into comprehensive managed care contracts.

The accounting is part of the problem. States can claim NEMT as a medical service, an administrative cost, or both. On the federal CMS-64 expense form for fiscal 2021, 90 percent of reported NEMT dollars were claimed as medical services and 10 percent as administration, and rides built into health plan capitation do not appear on those NEMT lines at all. For scale, CMS put total Medicaid spending at nearly $1 trillion in fiscal 2025. See who pays for NEMT for how the money reaches providers.

Companies and drivers

The Census Bureau’s County Business Patterns counts establishments with paid employees by industry. Its special needs transportation category (NAICS 485991) is the closest fit for NEMT companies. The 2023 file shows:

Two limits apply. Owner-operators with no payroll are not in these counts. And the category also covers senior, disability, and paratransit transportation that Medicaid does not pay for.

BLS wage data has no NEMT driver occupation, so the table uses the closest ones. These May 2025 estimates cover wage and salary workers only, not the self-employed.

The mean for shuttle drivers and chauffeurs was $19.17 an hour, above the median, because higher earners pull the average up. State and metro wages differ widely, and NEMT driver pay covers how operators set rates.

Fraud enforcement

NEMT draws steady enforcement attention. The HHS inspector general’s annual report on Medicaid Fraud Control Units, released in March 2026, counted 1,185 convictions in fiscal 2025, 856 of them for fraud.

• Criminal cases. NEMT providers accounted for 52 fraud convictions, tied with nurses for third place. Only personal care attendants (326 convictions) and non-residential mental health facilities (63) had more. Criminal recoveries in NEMT fraud cases came to about $8.39 million.
• Civil cases. NEMT ranked fifth for civil settlements and judgments, with 31, and about $21.94 million recovered.
• Audits. The inspector general added an NEMT audit series to its work plan on May 28, 2026. The audits check whether chosen states followed Medicaid payment rules for NEMT claims.

States can also close the door to new companies. Minnesota has stopped enrolling new NEMT companies based in the seven metro counties around Minneapolis and Saint Paul, and that freeze runs through January 27, 2027. See enrollment moratorium and NEMT fraud prevention.

Reading these numbers

Use these figures for direction, not as exact counts. Keep four limits in mind:

1. Ride days are not trips. One ride day can hold two or more legs.
2. Claims data misses some rides. Rides paid as an administrative expense, and some broker and plan encounters, never reach the national data.
3. Spending excludes managed care provider payments. The $2.6 billion figure is a floor.
4. Business and wage counts are partial. County Business Patterns skips owners with no payroll, and BLS skips the self-employed.

Market-size estimates that do not show their data or methods deserve extra caution. For planning a real company, local numbers matter more: your state’s program, the dialysis centers and senior communities near you, and the rates your payers publish. NEMT market research walks through how to gather them.

Tracking your own numbers

National statistics tell you the market exists. Your own trip and payment records tell you whether your company is working. HealthRide’s reports cover completed trips, on-time performance, driver activity, and a payer summary with what each payer was billed and what it still owes, with no report builder to set up. See reports.